fix(tokenize): correct capture group reference in website regex

The websites regex /[.](com|net|org|io|gov|edu|me)/g has only one
capture group for the TLD, but was incorrectly referencing $2.
This caused URLs to be corrupted with literal '$2' text.

Changed $2 to $1 to correctly reference the TLD capture group.

Author: IlyaShelestov

agents/src/tokenize/basic/sentence.ts

@@ -27,7 +27,7 @@ export const splitSentences = (
   }
 
   text = text.replaceAll(prefixes, '$1<prd>');
-  text = text.replaceAll(websites, '<prd>$2');
+  text = text.replaceAll(websites, '<prd>$1');
   text = text.replaceAll(new RegExp(`${digits.source}[.]${digits.source}`, 'g'), '$1<prd>$2');
   text = text.replaceAll(dots, (match) => '<prd>'.repeat(match.length));
   text = text.replaceAll('Ph.D.', 'Ph<prd>D<prd>');